Repair Worlds: Maintenance, Repair, and ICT for Development in Rural Steven J. Jackson Alex Pompe Gabriel Krieshok Department of Information Science School of Information School of Information Cornell University University of Michigan University of Michigan [email protected] [email protected] [email protected]

ABSTRACT novelty, growth, and progress as our starting point. This paper explores the nature and centrality of This perspective runs somewhat counter to the way we maintenance and repair (‘M&R’) work in the extension and usually approach questions of design and development, in sustainability of ICT infrastructure in the global South. both CSCW and the ICT and development (ICTD) fields. Drawing from pragmatist traditions in CSCW and the social Most accounts of ICT and development have emphasized sciences at large, we develop a concept of ‘repair worlds’ encounters between organizations, communities and intended to map the varieties and effects of such technologies at the front end, emphasizing pathways, maintenance and repair activities. Empirically, our analysis patterns, and problems that accompany the introduction of builds on ethnographic fieldwork into local practices of (new) technological forms to (new) geographic and cultural maintenance and repair that have accompanied and communities. ICT-related development statistics and supported the extension of mobile phone and computing studies have long reflected this bent. We keep track of infrastructure in the of northeastern figures like ICT penetration, adoption, and diffusion, but Namibia. rarely breakdown or abandonment. We capture statistics on cell phones purchased, but rarely cell phones discarded. We Author Keywords study how communities embrace or adapt to new Maintenance, repair, infrastructure, theory, ethnography, technologies, but less often how they organize to sustain, development, ICTD. manage, repurpose, or simply live with what they have. (As ACM Classification Keywords a thought experiment, consider how our statistical and H5.m. Information interfaces and presentation (e.g., HCI): design universes might shift if we imagined a Miscellaneous. fundamentally contracting technological world, and what we measured was technological breakdown, withdrawal and General Terms decline – a recessionary rather than a progressive, Management, Theory. informatics?).

That we don’t (usually) live in such a world may have less 1. INTRODUCTION to do with any inner logic of economic or technological This paper explores the centrality of maintenance and repair advance than with the remarkable and diverse human work in the ongoing project of building robust, appropriate, capacity for repair: the ability to make broken and breaking and sustainable ICT systems in countries of the global south systems work, and to sustain what we have for at least a – a project to which CSCW has much to contribute. It also little while longer. This sensibility has long animated an argues that CSCW has much to (re)learn in this process: important branch of American social thought, from John about the nature and variability of socio-technical activity Dewey’s insistence on disruption as site and instigation to (especially as played out over cultural and infrastructural thought [1], to the breaching experiments of difference); the work required to sustain and adapt systems ethnomethodologists like Garfinkel [2], to later Chicago over time (including beyond initial moments of design and School sociologists’ concerns with infrastructure, adoption); and the distinctively different worlds of design articulation, and invisible work [3-5]. Social order, such and practice that appear to us when we take erosion, work insists, is not the natural or inevitable outcome of breakdown, and decay (‘broken world thinking’) rather than ‘structure,’ but rather a remarkable and often fragile accomplishment. Social things, the pragmatists remind us, take work to maintain. And so, too, does technology. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are In the paper that follows, we explore maintenance and not made or distributed for profit or commercial advantage and that copies repair as distinctive sites and categories of CSCW work, bear this notice and the full citation on the first page. To copy otherwise, with particular focus on ICT and development efforts in the or republish, to post on servers or to redistribute to lists, requires prior global South. Drawing on original ethnographic fieldwork specific permission and/or a fee. CSCW’12, February 11–15, 2012, Seattle, Washington, USA. conducted during summer 2010 in the Kavango region of Copyright 2012 ACM 978-1-4503-0267-8/11/05....$10.00. Northeastern Namibia, we offer the beginnings of a thick

1 description of the distinctive repair worlds that characterize, ‘world,’ whose function and variety is central to the nature shape, and sustain the emerging ICT ecologies of the and overall success of ICT development in the region. region. In the process, we open up new sites and themes for To this end, we offer ethnographic data drawn from CSCW scholarship – many of which take us back to fieldwork conducted by two members of the project team challenges and problems foundational to the field. between May and July 2010 around ICT and educational 2. LITERATURE AND METHODS development efforts in the Kavango region of northeastern CSCW has a rich albeit uneven history of attention to Namibia. This included 34 semi-structured interviews with maintenance and repair concerns. In its design guise, such regional technology providers, managers, technicians, and themes are often implicated in CSCW work attending to users, conducted in a combination of English and ‘downstream’ tensions in the relationship between design Rukwangali (the leading regional language) and focusing and use, leading to ethnographic descriptions and/or on patterns, practices, tensions, and barriers to effective and programmatic calls for tailoring, end user involvement, and sustainable ICT adoption; and field observations at 14 more iterative or ecological approaches to the design different rural and urban sites in which primarily problematic [6-8]. A related body of work has explored educational technologies (usually computer labs) were instances of breakdown and non-adoption confronting the being deployed. Participants ranged in age from 20 to 55 deployment of CSCW artifacts and systems, including years old, were evenly split between men and women, and through contradictions embedded in the relationship were recruited according to their involvement as technology between technical and organizational form [9-11]. users, providers, or promoters in the region. Interviews and fieldnotes were translated (where necessary), transcribed, Our own approach, building from pragmatist traditions, is and coded by all three members of the project team to treat such instances of breakdown and failure as fully according to the principles of grounded theory analysis [18- ordinary, and ask instead (with genuine curiosity and 19]; early rounds of coding thus came to guide subsequent wonder) how it is that any sociotechnical arrangement can data collection efforts, including more focused attention be made to work and persist through time. Here we connect (per grounded theory principles of theoretical sampling) to to a third strain of CSCW work emphasizing the diverse the dynamics of maintenance and repair that emerged as an and subtle forms of work that go into the ongoing operation early and central theme of the fieldwork. This initial of artifacts and systems as ‘infrastructure’ (vis-à-vis given fieldwork has since been supplemented through ongoing forms of collective activity and variously situated social interactions with a team of local partners, who continue to groups). Star and Strauss have analyzed forms of conduct occasional interviews and supply materials relevant articulation work, much of it invisible, supporting the to the themes of the study. More broadly, our findings draw introduction and maintenance of collectively shared on one of the author’s prior experience as an IT and artifacts and systems [3]. In their 1994 CSCW study of the education volunteer in the Kavango region between WORM community system, Star and Ruhleder advance an November 2006 and January 2009. influential definition of infrastructure, emphasizing the embeddedness of infrastructure (in social arrangements and 3. REPAIR WORLDS technical form); its tendency to remain invisible (except at For Strauss and fellow travelers in Chicago School moments of breakdown); its connection to membership in sociology traditions, social worlds display a number of specific communities of practice; its relationship to an common properties. In each world, at least one primary installed base (from which it inherits both strengths and activity (along with related clusters of activity) is strikingly limits); and its tendency to change in local and modular evident: e.g., climbing mountains, researching, and increments – in part through the resistances and collecting. There are sites where activities occur, breakdowns it occasions and encounters [4]. Subsequent implicating place, landscape, and a shared built CSCW scholarship has built on this definition, further environment. Technology (understood broadly as inherited developing the active [12-13], relational [14], agonistic, or innovative modes of carrying out the social world’s [15-16] and dynamic [16-17] dimensions of Star and activities) is always involved. And while social worlds may Ruhleder’s original formulation. begin from temporary and provisional divisions of labor, at some point organizations typically evolve to structure and Our work brings a parallel sensibility to the analysis of ICT manage key aspects of the world’s activities [20]. and development efforts in rural Namibia, with special attention to the distinctive problems and solutions Each of these broad categories can be further specified surrounding sustainability in this space. The vehicle for our through a series of subprocesses, some of which may discussion is a distinctive notion of repair worlds, drawn constitute distinctive and legitimate ‘sub-worlds.’ These from the same social worlds tradition (discussed below) that may include partially autonomous activities dedicated to the has informed and inspired much CSCW scholarship around finding, funding, protecting, and competition for sites; infrastructure. We argue, in short, that ICT maintenance innovation, manufacturing, marketing, and training in and repair activities in the Kavango region of Northeastern technical skills; or the building, extending, defending, and Namibia constitute a distinctive and complex kind of repurposing of organizations. The discovery and study of

2 such sub-processes and of their relationships, including people and things, opening up both dimensions of human conflictual ones, are essential features of social worlds signification embedded in the world of things, and the research. possibility of strategic ‘switching’ between human- and object-centered approaches to collective goals (this Social worlds are also subject to complex processes of principle is nicely captured in Bruno Latour’s example of boundary formation, both internal and external. Externally, the sleeping policeman [22], or Lawrence Lessig’s principle relations and intersections between adjacent worlds may be of code [23]). sites of tension, innovation, and strategic action. Internally, social worlds may follow complex patterns of Finally, social world analyses have been historically good segmentation, giving rise to subworlds that will themselves at calling attention to the kinds of support work, people, and exhibit some of the dynamics (including boundary artifacts that underlie the actual practice of distributed dynamics) of worlds of a more general type. At such collective activity. From invisible work [3] to infrastructure junctures, a number of crucial questions arise. Who has the [4,5], interactionist traditions have drawn new attention to power to speak, decide, claim membership, and represent people, practices, and artifacts operating beyond or beneath on the part of that world and its inhabitants? How are the margins of more formally recognized systems. In the newcomers socialized, and what is the range of allowable process they have opened a space for precisely the sort of relationships (from core to more ‘orbital’ and overlapping maintenance and repair concerns pursued here. In Star and memberships)? How do social worlds – and the more Ruhleder’s influential account cited above, for example, formalized instantiations like arenas and organizations that infrastructure is defined as a sort of social and material grow from them – form and change, and how do they relate substrate or scaffolding to action that remains “invisible to larger processes of social transformation (economic, until breakdown” – at which point underlying processes are cultural, political, etc.)? made visible, and new repertoires of action may kick in [4]. Ethnomethodologically inspired traditions of CSCW A leading example of these ideas in action is provided by scholarship have relied on this same basic insight: for Howard Becker’s Art Worlds [21], which approaches the example, seminal studies by Lucy Suchman, Julian Orr and making of art as a form of materially and collectively others which follow (not by accident) the experience of organized activity. To succeed as art, argues Becker, Xerox copier maintenance and repair workers to produce objects depend on a wide and heterogeneous list of new insights around planning, design, and the nature of components: ideas and possibly inspiration; but also forms technically-mediated collective action [24, 25]. And of physical execution; materials (canvases and paint growing efforts by CSCW and HCI researchers to brushes, cameras, etc.); support work (from editing and foreground sustainability as a design value offer new ticket sales to the bringing of coffee); an audience, opportunities to push questions of maintenance and repair readership, or other network of response; a rationale or back to the center of the field [26-27]. In the sections that other (collective) justification of value; networks of follow we build from social world and interactionist training; and certain conditions of civic order which provide principles to describe the repair worlds of the Kavango a stability of expectations against which the basic region in Northeastern Namibia. operations of the world can proceed. In stable and coherent worlds, these components will reinforce and overlap – for 4. ICT, MAINTENANCE AND REPAIR IN THE example, in the way aspiring artists and actors constitute a KAVANGO significant portion of the audience for experimental or In the past five years Namibia has seen an enormous growth emerging art forms. They also generate constraints and of technological availability in both cellular technology and conventions that will shape and regulate the form and networked computing resources. Mobile phone ownership process of art – for example, conventions around symphony rates currently stand at over 50% of the population [28]. length and orchestra size, which shape and limit the nature While low by global standards, Namibia’s IT infrastructure of creative work in this space. remains among the most advanced in sub-Saharan Africa. And investment continues to grow. In 2008, Namibia’s For IT and development scholars and the wider CSCW largest cellular provider, MTC, spent an additional NAD76 field, such perspectives carry important advantages. To million (~US$10.1 million) in fiber optic backbone, begin, in contrast to structuralist traditions of social bringing the company’s total investments from 1995 thought, social world analyses start from an explicit through 2008 to NAD1.6 billion (~US$213 million) [29]. concern with social change – a significant advantage in the technology and development context – which nevertheless But such overall statistics mask significant regional takes instances of stasis and consistency as equally central variations. As with other national numbers, technology and phenomena of interest (and requiring explanation in any wealth indicators in the country are dominated by the careful analytic account). Second, in its gathering of capital city of , which continues to function as the ‘human’ and ‘technical’ elements under the undifferentiated reference point for questions of availability, pricing, and notion of worlds, social world analyses offer us tools for access. Because of this, nation-wide technology policies are thinking even-handedly across the interlinked worlds of often detached from the realities of the poorest regions,

3 such as the Kavango, where development energy and Training is obtained by the owners outside of the Kavango dollars are most needed. Reflecting apartheid policies in region, and in many cases outside of Namibia (of the three place before national independence in the 1990s, Namibia formal technology repair businesses currently found in is characterized by a Gini coefficient of 74.3, representing , two of the chief proprietors were educated in South the largest national wealth disparity in the world [30]. This Africa, and one in Egypt). The shops’ supply chains follow gap is largely reproduced in the balance of technology similar lines: two receive their parts from South Africa and resources between Windhoek and the Kavango, and again the third through an Egyptian supplier. Repairs at these between the regional capital of Rundu and the towns, shops usually consist of replacing failed hardware villages, and rural homesteads that surround it. components, re-installing software, and removing computer viruses (a nearly universal feature of the Namibian To redress this gap, the Namibian government and foreign computing landscape). donors have begun heavy regional investments in computing infrastructure, centered primarily in schools. An even larger and more varied market has developed Since 2006, the national government’s Tech/NA! program – around cell phone repair. By 2010, no fewer than ten such an ambitious and country-wide school computerization shops had opened in Rundu, with an additional three to be initiative – has been operative in the Kavango. As of July found in the smaller regional towns of Nkurenkuru and 2010, Tech/NA! labs and training centers had been set up at Divundu. Most of these operate out of single room roadside the Rundu College of Education, two Teacher Resource stalls, in some cases converted shipping containers. Others Centers (TRCs), and several senior secondary schools. This have attached themselves to more established and has been accompanied by a series of parallel volunteer and apparently unrelated businesses, such as bars, clothing donor-driven initiatives, including Computers for Kavango, shops, or our favorite, Cell and Leather World. Shop a project started by U.S. Peace Corps volunteers that uses owners are uniformly male but show a significant diversity donations raised through Facebook to purchase rebuilt of backgrounds. Many started as apprentices in other phone computer systems that are then shipped to Namibia and set repair operations (often in the national capital of Windhoek) up in smaller regional schools. The net result of these and before opening their own businesses in the Kavango. Others other initiatives has been a steep increase in the number and were educated abroad, with one individual having obtained availability of computers in the region: from an estimated an MBA degree in his native Pakistan. While some have fewer than 50 in 2006, to over 500 by May of 2010. formal IT training (e.g. at universities and polytechnics in Regional cell phone use and coverage has expanded even the capital), the vast majority – reflecting patterns in other more dramatically, with population network coverage local repair cultures (e.g. auto mechanics, tire repair shops) expanding from less than 50% to over 90%. – have acquired their skills, networks, and reputations through apprentice relations. The most common activities Nevertheless, as recently as 2007, there existed few formal engaged in by these workers are swapping batteries, operations in the Kavango region dedicated to ICT unlocking phones, re-flashing operating systems, and maintenance and repair. Users with broken mobile or replacing cracked screens. computing devices faced the prospect of working with locally available parts recycled from old and broken At the opposite end of the spectrum from these semi- machines; a trip to the national capital of Windhoek, ten formalized operations lies the Rundu repair market. This hours away; or a visit to the town’s informal repair market open-air market is located at the center of town and houses (described below). General computing equipment (memory a multitude of informal and small-scale enterprises grouped sticks, printer cartridges, blank CDs, etc.) was available in in a bazaar-like arrangement. While nominally a trading limited and sporadic supply, and only in the regional capital center, the open market functions as a defining social of Rundu. In 2008, the two national cellular carriers, MTC institution of the Kavango in general. It serves as a daily and CellOne (now Leo), opened their first formal sales stop for many of Rundu’s residents, whether to eat food, operations in Rundu; however, these remained focused on buy fabric, sell a chicken, or watch a soccer match. Over billing, phone, and SIM card sales, and offered no repair ten individual shops/stalls in the market deal in technology services beyond simple replacement. (both computer and cell phone); no fewer than five of these are dedicated to repair. The workers here offer skills and 4.1 Private Sector Sites services built up over years of experience in this By 2009, the local repair ecology had begun to change. environment, dominated above all by an ad hoc capacity for Three new ‘tech shops’ had opened, combining computer tinkering. This tinkering culture is identifiable by three sales and repair operations. A representative example of main characteristics. First, there is a tremendous amount of this shop type was opened by ‘Tom,’ a hulking Afrikaaner reuse and recycling of parts extracted from broken systems, from South Africa. To northern audiences, Tom’s shop will rather than the purchase of new replacement components (it look much like a family-owned small business: the shop is hard indeed to find new parts in the Rundu market). occupies a permanent and prominent location at the center Second, these repair shops operate largely beneath the of town, the majority of the staffing is done by Tom's expertise of a single person, occasionally supported by family, and the owner himself completes all repair tasks. apprentices (usually family or at least ethnically related).

4

Lastly, the cost of procuring repairs at the market is lower was a place where teachers could make photocopies, attend than any other local option. This makes the open market workshops, and check out library books. However, in recent notably cost-accessible, and thus the single busiest site by years the main resource utilized at the TRC has been its volume of maintenance and repair work in the region. technology. Beyond its large and functional computer lab and reliable broadband connection, the TRC also houses two IT repair experts who serve as de-facto doctors for all school computer illnesses in the Kavango. Over our three months of observation, an average of three computers a day were brought in and repaired at the Rundu TRC. The Rundu TRC’s success as a repair center depends on the on-site knowledge of its staff and its regionally unique combination of resources. The TRC has become a drop-off point for old computers as the staff has the knowledge to extract working components from non-working systems. When this stockpile of parts does not fill the need for a given repair, the TRC staff act as liaisons to the private sector shops, such as Tom’s. In addition, the TRC’s generally solid broadband connection provides the rare ability for large software updates to be downloaded, in particular updates to operating systems and anti-virus software. The same Internet connection allows TRC staff to search for additional information when diagnosing and resolving unfamiliar maintenance and repair issues. As the typical computer lab in a Kavango classroom involves a wide array of hardware and software components, including sometimes spotty second-hand donations from the United States or Europe, such situations arise frequently. 4.3 Rundu Institute of Computing A second quasi-public repair operation can be found in the Rundu Institute of Computing (RIC), located in a re- purposed concrete structure near the center of town between the new Chinese built shopping mall and a series of road- Figure 1: Repair worker in the Rundu open market, side stalls made from wood and discarded sheet metal. The surrounded by discarded electronics to be salvaged for center provides a variety of technology related services, but working parts. its most profitable are the repair of computers and cell 4.2 Public Sector Operations phones. The shop and its owner, Andreas, enjoy a reputation as the best place to take any cell phone that is In addition to these private sites, a number of regional broken. Andreas performs most of the computer repairs repair operations have grown up to serve the growing himself; his former students turned employees perform the market in government or donor-supported computing bulk of cell phone repairs. Like the stalls in the open initiatives. Reflecting the older but slower spread of market, the Rundu Institute offers prices lower than those at computing infrastructure in the region, the vast majority of shops like Tom’s. Also like the market, most components computers in the Kavango today remain the property of are repurposed or sourced locally (albeit with a higher government ministries, and in particular the Ministry of reputation for quality and provenance: the RIC is not Education. Official maintenance policies (explored in known to participate in the growing black market for stolen greater detail in [31]) state that all computers purchased by phones). In addition to the expertise of Andreas and his the government must be serviced by government-owned or staff, the RIC depends on a now-considerable stockpile of approved operations. This poses significant challenges, for components drawn from the dead and unfixable machines the projects themselves and for local repair worlds, as and handsets periodically brought to its door. The business discussed below. has been in operation in one form or another for 7 years, The primary regional computer repair center operated by and at its current location since 2005. the Namibian government is located in the Rundu Teacher Resource Center (TRC), a center administered jointly by the Ministry of Education and the National Institute for Education Development (NIED). Traditionally, the TRC

5

software settings. The pervasive practice of SIM card swapping amongst members of families and social groups extends the use of a single handset (as widely celebrated in the IT and development literature), but also leads to the phone often becoming locked by mistake; for example, when SIM card passwords are entered incorrectly. However, with the use of special hardware components, the phone’s software can be effectively hacked when connected to a PC. This solution depends on distinctly non-local ties, however; the necessary hardware components are not sold commercially in Namibia, and have been sourced (in this instance) from Russia. In the absence of local infrastructure, Andreas is similarly dependent on remote partners for learning and adaptation in the fast-moving phone market; as he explains, “I am still learning many things, as the cell phones are changing quick. I meet partners on the Internet, on forums, who help me with my ideas… I have one partner from South Korea; (he) gives advice and has taught me many things. I tell my students that your partners are the most important thing in this business.” Figure 2: An apprentice at the Rundu Institute of Computing repairing a broken cell phone. 5.2 Disc Sharing 5. CHALLENGES AND INNOVATIONS A second innovative practice uncovered in the course of our fieldwork applies to the maintenance and repair of The variety of sites detailed above contributes to a computers in schools. Most computers arrive at schools as significant range and heterogeneity in the repair worlds second hand donations without any software discs. Key (and sub-worlds) of the Kavango region, each of which can barriers to on-site repair are limited repair knowledge of be arrayed differently along axes of formality, quality, school staff, a lack of reliable Internet connections to profitability, and institutional and social sectors served. diagnose problems, and the challenge of transporting This has produced a relatively varied and robust landscape broken machines over distance to the Rundu TRC. Thus, of maintenance and repair in the region, with important the most common software repair technique is to reinstall (albeit overlooked) contributions to the spatial and social the computer’s operating system. It is common for a extension and sustainability of ICT infrastructure in the computer to be given a fresh install, become infected with region. It has also produced some distinctive challenges and viruses (spread with USB memory sticks), and then be innovations, shared in story form below. given a fresh OS installation within 2-3 months. Most repair staff are capable of starting from scratch with a system with 5.1 Cell Phone Hacking the use of a system disc. This represents the most basic One of the most common cell repairs performed at the level of repair knowledge, but is also the practical reality Rundu Institute of Computing is the unlocking or re- for most computers in the region. flashing of cell phone operating systems. The institute’s The lack of broadband connections limits the availability of owner, Andreas, initially repaired only computers but large operating system updates. It is not uncommon to find quickly realized that the cell phone market was much larger multiple copies of the same out of date version of Linux or and profitable. He used his knowledge of electronics and Windows being used at many schools in the region. These software systems to teach himself basic cell phone repairs. discs (and their pirated software keys) trace interesting While he did not yet own a computer, he would visit resource-sharing routes amongst the Kavango’s repair Rundu’s first dial-up Internet café in a local photocopy world. We found that discs follow the paths of students as shop in order to research the subject. This allowed him to they graduate, eventually reaching as many as five schools. learn many aspects of cell phone hardware repair at a time One lab administrator near a border town with coinciding with their mass adoption in the Kavango. Since recalls the following example: 2008 Andreas has had access to a personal Internet connection and his cell phone repair abilities have grown “I was only taught how to install. I use this disc every week enormously. (copy of operating system disc). I copied it at my first school Himarwa when I was in grade 12 from a teacher Andreas has used his Internet connection to circumvent a there. I moved here and I knew that I would need it… Yes, I key problem of cell phone software repair: most phones do also share with others; other teachers at schools down the not allow easy access to manipulate or reset foundational

6 road. They need it to fix their machines.” a strong relationship with several of his students who later joined him at the institute upon its formation seven years The adaptation and use of social networks is an innovative ago. Andreas teaches a year-long course in electronics and and prevalent tool used by actors in the Kavango’s repair computer repair that he developed himself. world. Just as traditional tools are often shared amongst an extended family, operating system discs find their usage Typically two to three youth under the age of 25 will multiplied through sharing. complete this training each year, and the majority of them then work at the Institute itself. After an average of two to 5.3 Computer Labs as Business Models three years, many of the apprentices will move on and start A key challenge for the maintenance and repair of their own repair-centered business. (Several of the semi- computers at schools in the Kavango is providing the formalized cell phone repair shops attached to local bars are upkeep costs for their labs. Few projects allocate any funds former students of Andreas.) Beyond serving the cell phone for ongoing maintenance or eventual upgrades of maintenance and repair needs of the Kavango region, the computers, instead defining success as the computers Rundu Institute for Computing also serves as a transit point working when they were installed at the schools. This is between formal expertise (as represented by Andreas’ especially negligent given that many computers are second UNAM degree and experience at Telecom Namibia) and hand or donations and are living a second life after having less formalized forms of work in the regional repair been discarded by their initial owners. The limited capital economy. But this between space also poses challenges. To of schools is thus a major challenge for maintaining labs of begin, the semi-formal nature of the Institute limits aging hardware. The more rural a school, the less breathing Andreas’ ability to attract outside lending, and thus to room in its budget, and the more expensive it is to obtain expand and upgrade his repair operations by hiring and transport replacement components. additional UNAM graduates. This prevents Andreas from off-loading some of his administrative responsibilities to To deal with this problem, the Computers for Kavango others, as he’d like to do; as he explains, initiative has introduced a model whereby unemployed youth in communities are trained in basic maintenance and “I prefer (to work) in the research, in development. The repair of computers. In exchange for maintaining the labs, basement of this business is where I belong. I am not good the youth are allowed to make use of the computers for at bringing in people and handling administration matters, income generating small to medium-sized business but (my) time is spent there now.” enterprises, ranging from photo printing to basic typing and The Institute has also struggled to build productive document services (including personal resumes). Lab partnerships with the many formal volunteer and administrators are also allowed to offer basic computer international development agency projects that have grown classes to the community at large during school off-hours in in the region in recent years. By overlooking repair exchange for a small tuition. Several of the youth operations such as Andreas’, projects in effect lock them administrators we interviewed had high enrollment in their out (to the detriment of both parties). The reality of the courses and took great pride in their role as technology repair and maintenance landscape currently found in the liaisons for the rest of the community. Kavango is that there is no shortage of broken computers in 5.4 Biography and Career Trajectories need of timely and cost effective repair. However, due to its insufficiently formalized status, one of the most skilled To be robust and sustainable, social worlds must also: a) technology repair institutions in the region is unable to offer their members viable paths to learning and access this valuable market. advancement; and b) provide mechanisms for the social reproduction of the world itself. This raises important 5.4.2 “Wesley” questions of biography and career trajectory (and restores to Social reproduction through career evolution is also individual actors a degree of specificity and agency lost apparent in the following account of repair practice in the under more structuralist analytic traditions). In the limited smaller town of Nkurenkuru, an hour to the west of Rundu. space available, we offer two brief biographical sketches of A teacher at a local high school was quick to tell us the key roles and actors in our immediate study sites. repair work that his son Wesley performs while at home 5.4.1 “Andreas” during school breaks. During most of the year Wesley studies computing technologies at the University of A defining characteristic of the Rundu Institute of Namibia, the country’s flagship institution located in the Computing is its use of teaching and apprentice style capital of Windhoek. Outside of school, he finds many relationships. Andreas himself brings a formal education in informal maintenance and repair jobs at other informal computer and electronic technologies from the University businesses around Nkurenkuru. His father shared with us: of Namibia. Before opening the institute, he worked in the formal industry sector for Telecom Namibia, and then for “My boy is coming here now to the market to work on two three years as a school teacher of computer studies in the computers just there. They know and wait for him to return Rundu area. In his time as an educator, Andreas developed from school. They call him on his cell phone. I saw him the

7 other day with some very nice shoes and trousers I did not capital given high-level administrative passwords and give to him. They came from money he earned finding work settings. In additional to the logistical logjams created, such on these computers.” excessively rigid repair policies marginalize local repair actors and prevent them from evolving new technical skills. In dialogues with other community members, it became To solve at least the first problem, the local TRC staff has apparent that many technology users around Nkurenkuru hit upon the solution of simply re-flashing the client by knew and depended on Wesley, and planned their computer booting with a USB memory stick, restoring all software repairs around the school holidays when he would return settings to their initial state and usually correcting the home. While Wesley's formal expertise and standing is problem (though at the sacrifice of all stored files). To identifiable through his status as a student at university, the avoid the loss of computers through trips to the capital, the majority of his practical repair expertise has been developed TRC staff now distributes these USB sticks with usage in the informal setting. Wesley's experience while still a instructions to all affected Tech/NA! labs. student not only augments his formal education; it provides him and his family a greater source of income to help 6.2 Problems of Lockout and Control absorb the costs of such an education. In addition, he When computers are brought to the TRC for repair, it is provides a low-cost repair solution that maintains the quite common for them to contain many files that have been technology landscape of a smaller town such as copied to the machines by users. It is common practice for Nkurenkuru. the staff to back up local files before the installation of a 6. CHALLENGES new hard drive or operating system. However, a large percentage of these files may consist of music collections or If the above stories offer evidence of promising and personal photographs, which the TRC staff deems non- innovative workarounds to address the distinctive educational. They thus do not back up these music files, and challenges of maintenance and repair in the Kavango, other speak to the school staff on the dangers of spreading problems have proven more intractable. Our fieldwork computer viruses through USB sticks by sharing music. In uncovered other instances in which the range and addition, the TRC may discourage “inappropriate” use by heterogeneity of repair worlds in the region and their disabling the sound card and other system settings (which interface with other institutional realities have constituted can only be restored through an administrator account, liabilities rather than assets. which teachers and students are forbidden from attaining). 6.1 Policy Barriers On the surface this appears as a pragmatic response to the In several instances, we found significant mismatches real problem of users inadvertently breaking and infecting between the formal maintenance and repair policies of computers. But such control mechanisms also shape and computers for education initiatives in the region and limit the adoption of computers in schools. While the use of resources and realities of local repair worlds. The Tech/NA! computers for educational purposes is a worthy objective of program, for example, follows a centralized repair model in computerization efforts in the Kavango, the most powerful which all maintenance and repair operations are to be local incentives to engage these efforts often lie in other conducted at the newly created National Education interests and activities (e.g. music and entertainment). Our Technology Service and Support Centre (NETSS Centre) own fieldwork and experience suggests that the students located in Windhoek. Strictly followed (as it sometimes is), and teachers leading adoption in the Tech/NA! and this policy requires all machines requiring attention Computers for Kavango labs are precisely those who have (including for simple software upgrades and other routine become most proficient at using (and sometimes hacking) maintenance) to undergo the expensive and somewhat risky the systems for entertainment purposes. Along with the trip to Windhoek via the national post system. Machines restrictive repair policies noted above, this tension between then join a repair queue stretching several weeks to several freedom, innovation, and control constitutes the single months before being returned (or not) to their original largest tension attending use of the labs today. location. In the course of our fieldwork we encountered 7. DISCUSSION: IMPLICATIONS FOR POLICY AND several labs that had been effectively shuttered by this DESIGN process, with little immediate indication of restarting. This The concept of repair worlds advanced here and the more was despite the presence of a similarly government-owned specific case of repair worlds in the Kavango region resource in the Rundu TRC (to say nothing of the other produce few immediate design implications of the sort sites noted above) fully capable of providing a large common to CSCW scholarship (and questioned in [32]). percentage of the necessary repairs. But they do carry important lessons of both a practical and This conflict brings a power dynamic embedded in wider theoretical sort. infrastructure development to the forefront of repair The first and most obvious of these concerns the need for practices in the Kavango Tech/Na! labs. The Tech/NA! better bridges between formal development project systems have their software installed and configured in initiatives and the local repair worlds of the regions in Windhoek prior to their deployment, with only staff in the

8 which these projects are situated. The absence of this Finally, as alluded to above, the cases offered here connection (as we see in each of the TechNA and challenge the presumed separation between the design of IT Computers for Kavango cases above) represents a double artifacts and the design of IT policies that has tended to loss: for the projects, which see efficacy severely limit CSCW work to date. In the repair worlds of the undermined by the inability to effect timely and locally Kavango region, this line is difficult to draw. From the appropriate repairs; and for the repair worlds themselves, standpoint of programmatic interventions like TechNA! or which lose an important chance for support, growth, and the Computers for Kavango, the design of artifacts and the upgrading of technological skills, which turn out once again design of policy are better imagined as an interrelated under the old law of uneven development to be (re)centered whole whose integration or appropriate relation will go a on the capital [33]. long way towards determining success of the larger enterprise. By holding too faithfully to this divide, CSCW Such efforts to lock out and control use through repair- may be neglecting important and necessary sites of unfriendly policies and design follows from a long and intervention and thereby falling once again into what a unfortunate history of control through design in the ICTD respondent once identified to us as an “edifice complex” – and wider international development context. From electric the stubborn tendency to over-privilege artifacts or systems lighting kits [34] to telecenters [35] to hundred-dollar to the neglect of the supporting contexts in which their laptops, designers working in the development space have coherence and efficacy is sustained. This too is a basic long shown a heightened concern with damaging or lesson of social worlds scholarship. inappropriate use – and corresponding efforts to coax downstream users into ‘proper’ modes of engagement 8.CONCLUSION through design choices that lock out would-be tinkerers, As the sometimes rapid, sometimes halting spread of ICT modifiers, and adaptors. Sometimes this is for frankly networks, devices, and practices has unfolded over the past paternalistic reasons – for example, perennial donor hand- ten-plus years in Sub-Saharan Africa and elsewhere, the wringing around the fact that development recipients may kinds of repair worlds described here have come to play an be using the resources of donor-supplied ICT resources for increasingly central role. Recognizing, analyzing and “trivial” reasons: checking football scores rather than crop supporting this work promises important practical and prices; writing love notes rather than resumes. Regardless theoretical gains. Repair worlds may constitute interesting of inspiration, by limiting or short-circuiting the and overlooked sites of local innovation and expertise. development of local repair worlds, such efforts to manage They may support important kinds of extension and use through design and policy seem likely to undermine, tailoring, stretching and repurposing extant infrastructure to rather than support, long-term goals of sustainability. new geographic and social communities. They may remediate traditional relations of dependency and social Our third and related point has to do with the distinctive exclusion, sometimes amplifying these and sometimes opportunities for learning that breakdown, maintenance, ameliorating them. And by supporting and augmenting and repair provide – within but also well beyond the locally appropriate infrastructures, they may support new international development context studied here. Our forms of distributed collective work that can further broad favorite example of this comes from parallel fieldwork in developmental goals of economic, social, and Madagascar, where we encountered Mr. Rakoto, a former environmental sustainability and justice. electrician and now owner-proprietor of a phone and computing repair shop in the country’s eastern Mananjary Acknowledgements: The authors wish to acknowledge the region. Here’s how Rakoto describes his introduction to his support and contributions of Flay Julius, Disho Muruti, current line of work: Benjamin Sigrin, and Alex Voltiviski; without their help this work would not have been possible. “When cellphones first came to Madagascar, I was curious to see how they functioned, and how they might break. So I 9. REFERENCES bought one, threw it into the river here and then took it [1] Dewey, J. 1896. The Reflex Arc Concept in apart to see how the water and soil damaged it, and how it Psychology. Psychological Review, 3. 357-370. might be fixed.” [2] Garfinkel, H. 1967. Studies in Ethnomethodology, This represents, we think, a strikingly different way of Prentice Hall, Englewood Cliffs, NJ. approaching the problem of maintenance and repair than the [3] Star, S.L. and Strauss, A. 1999. Layers of Silence, one dominating northern imaginations of IT design and Arenas of Voice: The Ecology of Visible and Invisible sustainability, in international development and possibly Work. Computer-Supported Cooperative Work, 8. 9- other CSCW contexts. It constitutes a local version of what 30. we’d like to call “broken-world thinking”: a gestalt shift in [4] Star, S.L. and K. Ruhleder. 1994. Steps Towards and our ways of thinking about sociotechnical system Ecology of Infrastructure: Complex Problems in development that moves moments of maintenance and Design and Access for Large-Scale Collaborative repair, rather than just moments of design and adoption, to the heart of CSCW thinking and practice.

9

Systems. Proceedings CSCW 1994. New York: ACM. [19] Charmaz, K. 2006. Constructing Grounded Theory: A 253-264. Practical Guide Through Qualitative Analysis. Sage, [5] Star, S.L. 1999. The Ethnography of Infrastructure. London. American Behavioral Scientist, 43. 377-391. [20] Strauss, A. 1978. A Social Worlds Perspective. Studies [6] Trigg, R.H., Moran, T.P. and Halasz, F.G. 1987. in Symbolic Interaction, 1 (1). 119-128. Adaptability and Tailorability in NoteCards. [21] Becker, H.S. 1982. Art Worlds. University of INTERACT. Stuttgart. 723-728. California Press, Berkeley. [7] Nardi, B.A. 1993. A Small Matter of Programming: [22] Latour, B. 2000. When things strike back: A possible Perspectives on End-User Programming. MIT Press, contribution of 'science studies' to the social sciences. Boston. British Journal of Sociology, 51 (1). 107-123. [8] Dourish, P. and Edwards, W.K. 2000. A Tale of Two [23] Lessig, L. 1999. Code And Other Laws of Cyberspace. Toolkits: Relating Infrastructure and Use in Flexible Basic Books, New York. CSCW Toolkits. Computer-Supported Cooperative [24] Suchman, L. 1987. Plans and Situated Actions. Work 9(1). 33-51. Cambridge University Press, Cambridge. [9] Bowers, J. 1994. The Work to Make a Network Work: [25] Orr, J.E. 1996. Talking About Machines: An Studying CSCW in Action. Proceedings of CSCW Ethnography of a Modern Job. Cornell University 1994. Chapel Hill: ACM. 287-298. Press, Ithaca. [10] Grudin, J. 1988. Why CSCW Applications Fail: [26] DiSalvo, C., Sengers, P, and Brynjarsdottir, H. 2010. Problems in Design and Evaluation of Organization Mapping the Landscape of Sustainable HCI. Interfaces. Proceedings of CSCW 1988. Portland: Computer-Human Interaction. Atlanta: ACM. 1975- ACM. 85-93. 1984. [11] Orlikowski, W. 1992. Learning from Notes: [27] Grinter, R., et al. 2010. Computing at the Margins. Organizational Issues in Groupware Implementation. Georgia Institute of Technology, Atlanta. Proceedings of CSCW 1992. Toronto:ACM. 362-369. [28] Dutta, S. and I. Mia. 2010. The Global Information [12] Karasti, H. and Syrjanen, A-L. 2004. Artful Technology Report 2009-2010: ICT for Sustainability, Infrastructuring in Two Cases of Community PD. International Telecommunications Union. Participatory Design Conference (PDC). Toronto: CPSR. 20-30. [29] Telegeography. 2008. MTC Namibia Invests USD40m, Ends Reliance on Incumbent's Transport Network, in [13] Pipek, V. and Wulf, V. 2009. Infrastructuring: Towards Telegeography Research, PriMetrica. an Integrated Perspective on the Design and Use of Information Technology. Journal of the Association of [30] UNDP. 2010. Human Development Report 2010, Information Systems (JAIS), 10 (5). 448-473. United Nations Development Program. New York. [14] Star, S.L. and Lampland, M. Reckoning with [31] Jackson, S.J., Pompe, A., and Krieshok, G. 2011. Standards. In Lampland, M., and Star, S.L. ed. Things Fall Apart: Maintenance, Repair, and Standards and Their Stories: How Quantifying, Technology for Education Initiatives in Rural Namibia. Classifying, and Formalizing Practices Shape Proceedings of the iConference 2011. Seattle: ACM. Everyday Life. Cornell University Press. 3-24. 83-90. [15] Ribes, D. and Finholt, T. 2009. The Long Now of [32] Dourish, P. 2006. Implications for Design. in CHI Infrastructure: Articulating Tensions in Development. 2006. Montreal:ACM. 541-550. Journal of the Association of Information Systems 10 [33] Smith, N. 2008. Uneven Development: Nature, Capital, (5). 375-398. and the Production of Space (3rd ed). University of [16] Jackson, S.J., et al. 2007. Understanding Infrastructure: Georgia Press, Athens. History, Heuristics, and Cyberinfrastructure Policy. [34] Akrich, M. 1992. The De-Scription of Technical First Monday, 12 (6). Objects, W.E. Bijker and J. Law, ed. in Shaping [17] Edwards, P.N., Bowker, G.C., Jackson, S.J., and Technology/Building Society: Studies in Sociotechnical Williams, R. 2009. Introduction: Towards an Agenda Change. MIT Press: Cambridge. 205-224. for Infrastructure Studies. Journal of the Association of [35] McNamara, K. 2003. Information and Communication Information Systems (JAIS), 10 (5). 364-374. Technologies, Poverty, and Development: Learning [18] Strauss, A. and Corbin, J. 1998. Basics of Qualitative from Experience. World Bank,Washington DC. Research Techniques and Procedures for Developing Grounded Theory (2nd ed.). Sage, London.

1 0